For the past eight years, the PIs (M. R. Banaji and A.G. Greenwald) have collaborated on investigations if implicit social cognition. The main goals of the research program have been to demonstrate the operation of implicit social cognition, and to champion the view that theoretical, methodological, and application gains can result from investigations of processes that are ordinarily hidden from awareness. Past work focused on the stereotype domain, with published articles documenting the use of a variety of methods to reveal the operation of implicit stereotypes (Banaji and Blair, 1996; Banaji and Greenwald, 1995; Banaji and Hardin, 1996; Banaji, Hardin, and Rothman, 1993). A first goal of the proposed research is to extend investigations of implicit social cognition to the attitude and self-esteem domains identified by Greenwald and Banaji (1995). Greater attention to the constructs of attitude and self-esteem allows the development of a unified theory of implicit social cognition. Existing measures of automatic belief and attitude (especially ones based on cognitive priming) have been used effectively at some laboratories including our own, but have not been easy to transport across laboratories. It is also the case that existing methods (consisting of variations of priming techniques) have not yielded large effect sizes. Recently published work has produced a new measure (the Implicit Association Test, IAT; Greenwald, McGhee, and Schwartz, 1998) that forms the foundation of the present proposal. Several experiments conducted with the measure indicate that it can provide a reliable, efficient, and sensitive measure of individual differences in implicit social cognition. A second goal of the proposed research is to realize the potential of the IAT, method and to further develop it as a core technique for investigating implicit social cognition. This proposal consists of experiments that will further establish the generality and validity of the technique followed by three sections with a theoretical focus on prejudice and stigma, self-esteem and self-concept, and stereotype. Two additional goals, both occupying a less central position in this cycle of the research agenda, are mentioned. The IAT, it is to become a widely useable technique to study implicit social cognition more generally, will require psychometric attention. Therefore, a third goal of the proposed research is to secure the computerized procedure and scaling properties of the measure. Finally, it is a goal of the proposed research to attempt basic research with a clear eye toward the application of implicit attitudes, stereotypes, and self-esteem findings. The scope of the present grant restricts our ability to develop these ideas here. Hence, in the present proposal we will only indicate the path to such opportunities and preparation for future independent development.